Critical Role of the Fifth Domain of E-Cadherin for Heterophilic Adhesion with E7, But Not for Homophilic Adhesion
http://www.100md.com
免疫学杂志 2005年第14期
Abstract
The integrin E7 is expressed on intestinal intraepithelial T lymphocytes and CD8+ T lymphocytes in inflammatory lesions near epithelial cells. Adhesion between E7+ T and epithelial cells is mediated by the adhesive interaction of E7 and E-cadherin; this interaction plays a key role in the damage of target epithelia. To explore the structure-function relationship of the heterophilic adhesive interaction between E-cadherin and E7, we performed cell aggregation assays using L cells transfected with an extracellular domain-deletion mutant of E-cadherin. In homophilic adhesion assays, L cells transfected with wild-type or a domain 5-deficient mutant formed aggregates, whereas transfectants with domain 1-, 2-, 3-, or 4-deficient mutants did not. These results indicate that not only domain 1, but domains 2, 3, and 4 are involved in homophilic adhesion. When E7+ K562 cells were incubated with L cells expressing the wild type, 23% of the resulting cell aggregates consisted of E7+ K562 cells. In contrast, the binding of E7+ K562 cells to L cells expressing a domain 5-deficient mutant was significantly decreased, with E7+ K562 cells accounting for only 4% of the cell aggregates, while homophilic adhesion was completely preserved. These results suggest that domain 5 is involved in heterophilic adhesion with E7, but not in homophilic adhesion, leading to the hypothesis that the fifth domain of E-cadherin may play a critical role in the regulation of heterophilic adhesion to E7 and may be a potential target for treatments altering the adhesion of E7+ T cells to epithelial cells in inflammatory epithelial diseases.
Introduction
E-cadherin, a classic member of the cadherin superfamily, is expressed on epithelial cells and mediates Ca2+-dependent homophilic cell-cell adhesion (1, 2, 3). Classic cadherins contain five extracellular domains (ECs)2 of 110 aa each, a transmembrane domain, and a cytoplasmic domain. Structure-function analyses of the homophilic interactions of E-cadherin have largely focused on the NH2-terminal EC domain (EC1), which contains a highly conserved His-Ala-Val motif (4, 5, 6, 7). Indeed, protein fragments or peptides containing the His-Ala-Val sequence exerted limited effects on cell-cell adhesion (8, 9). Recently, crystallographic analysis has clearly demonstrated that conserved Trp in the EC1 domains of classical cadherins is critical for trans-interactions between E-cadherin molecules on different cells, serving as a strand dimer (10). However, several studies have suggested that ECs may be involved in cell adhesion in ways other than the role mediated by EC1. In human cancers, for example, E-cadherin gene mutations frequently occur in exons 7, 8, and/or 9 (corresponding to EC2 and EC3); these mutations are thought to result in the loss of the ability to undergo cell-cell adhesion (11, 12, 13, 14). A study on the binding properties of the soluble C-cadherin ectodomain suggested that EC1 was not sufficient for complete homophilic binding (15). Furthermore, Corada et al. (16, 17) demonstrated that mAbs directed against EC3-EC4 affected VE-cadherin adhesion in endothelial cells.
The heterophilic interaction of E-cadherin and integrin E7 has been previously documented (18, 19, 20, 21, 22, 23, 24). Integrin E7 is expressed selectively on intestinal intraepithelial T lymphocytes under physiological conditions (25). Accumulating evidence indicates that E7+ is induced on T lymphocytes in the epithelia of skin, lung, salivary, and lacrimal glands and synovial membranes during inflammation (26, 27, 28, 29, 30, 31), suggesting that this heterophilic interaction has a pathologic role. Since E7 may have an important role in the selective localization or retention of a unique population of T cells in a specific tissue, the adhesion between E7 and E-cadherin could be a potential target of therapeutic interventions for epithelial inflammation. Substitutions of a highly exposed and charged amino acid on mouse E-cadherin transfected into L cells demonstrated that Glu31 in EC1 is critical for binding with E7 (32). Taraszka et al. (33) also elucidated that the substitution of the corresponding Glu in EC1 of human E-cadherin-Fc fusion proteins abrogated binding with E7, confirming the previous observations. These studies clearly show that the specific heterophilic adhesion attributed to the exposed Glu31 in EC1 differs from homophilic adhesion. With the exception of EC1, however, the structures involved in heterophilic adhesion remain uncertain. To clarify the involvement of the EC domains in both heterophilic and homophilic adhesion, we performed cell aggregation assays using L cells transfected with specific domain-deleted E-cadherin mutations. The present report speculates on the characteristics of the E-cadherin domains involved in homophilic interactions and heterophilic interactions with E7.
Materials and Methods
Following the cell aggregation assays, the cell aggregates were collected and washed with PBS. For immunofluorescence staining, the aggregates were incubated with anti-E-cadherin Ab (HECD-1). Subsequently, the samples were incubated with Alexa Fluor 568-conjugated goat anti-mouse IgG Ab (Molecular Probes), followed by incubation with FITC-conjugated anti-E7 Ab (Beckman Coulter). The specimens were observed using a confocal laser-scanning microscope.
Results
To confirm whether the transfectants expressed the desired region of E-cadherin, RT-PCR was performed for EC1, EC2, EC3, EC4, and EC5. Fig. 2 shows that the parent L cells did not exhibit any products amplified by human E-cadherin and that all five ECs were expressed in wild-type transfectants. All clones transfected with domain-deletion mutations of E-cadherin exhibited the expected expression patterns: 1 lacked EC1, 2 lacked EC2, 3 lacked EC3, 4 lacked EC4, and 5 lacked EC5.
The protein expression of mutated E-cadherin was examined by Western blot analysis using the appropriate Abs (Fig. 3a). 4A2C7, raised against a recombinant protein corresponding to the cytoplasmic domain, detected all of the deletion mutants. G-10, an alternative E-cadherin Ab that cross-reacts with mouse form (manufacturer’s data), could not detect endogenous mouse E-cadherin in mock transfectants (data not shown). To clarify the binding epitopes recognized by the SHE78-7, HECD-1, and G-10 Abs, GST fusion proteins containing the EC domains were analyzed (Fig. 3b). SHE78-7 reacted with EC1, whereas HECD-1 reacted with EC2. G-10, raised against a recombinant protein corresponding to EC5, correctly detected EC5.
Discussion
In this study, we attempted to clarify the overall binding capability of E-cadherin for both homophilic adhesion and heterophilic adhesion to the integrin E7 through the generation of domain-deletion E-cadherin mutants and the transfection of these mutants into L cells to determine the roles of the individual domains in cell adhesion. EC1, EC2, EC3, and EC4 domain-deletion mutants lost their homophilic binding ability, suggesting that these EC domains are indispensable for homophilic adhesion. Substantial evidence suggests that the combination of cis-dimerization of two cadherin molecules on the same cell surface and trans-interactions between cadherin dimers on opposing cell surfaces maximizes homophilic adhesion. The widely accepted linear zipper model attributes the adhesive interfaces in the cis- and trans-interactions to EC1 (6, 7, 36). However, recent studies have introduced a new model for homophilic adhesion. Through the analysis of domain deletions in the Xenopus C-cadherin ectodomain using bead aggregation and cell adhesion assays, Clappuis-Flament et al. (15) demonstrated that the combination of at least three EC domains, such as EC1-EC2-EC3 or EC1-EC2-EC4, was required for trans-interaction and proposed an alternative model in which multiple ECs are required to achieve full adhesive capability. Several studies have provided data to support this model, suggesting the existence of multiple adhesive interfaces. For example, a mAb recognizing a potential epitope spanning the EC3–EC4 domains affected VE-cadherin adhesion in endothelial cells from umbilical veins (16). A mutant lacking the EC1 domain failed to exhibit homophilic adhesion, although the mutant could adhere to other mutants expressing wild-type E-cadherin (37). In contrast, a crystallographic analysis of C-cadherin supported the linear zipper model by showing that EC1 interacts with EC2 on other E-cadherin molecules on the same cell surface, leading to cis-dimerization (10). Our results showing that EC1- or EC2-deficient mutants failed to exhibit full homophilic adhesion are basically consistent with the above findings. However, mutants lacking either the EC3 or EC4 domains also failed to exhibit maximum adhesion. The fact that the present study used mammalian cadherin, whereas the previous study used Xenopus cadherin, may partially explain this discrepancy in findings. Another possible explanation may be that intracellular events were affected by the domain deletion in our aggregation assay, since the transmembrane and cytoplasmic domains were included in the domain-deletion human E-cadherin mutants. Compelling evidence suggests that the adhesive strength of cadherin is regulated by the transmembrane and cytoplasmic domains (3, 38, 39, 40, 41, 42). For example, the cytoplasmic tail of cadherin interacts with -catenin and p120, which link cadherin to the actin cytoskeleton through -catenin, and the regulation of the cadherin-catenin complex by diverse phosphorylation reactions influences the adhesive function of cadherin (43). In addition, the interaction of a motif in domain 4 of N-cadherin with the fibroblast growth factor receptor is required for neurite outgrowth (44), suggesting that EC4 may be involved in a cell signaling pathway in which fibroblast growth factor receptor controls the gene transcription and adhesive activity of cadherin via Snail, resulting in a loss of cell-cell adhesion (43). To the best of our knowledge, little information is available about the involvement of the membrane proximal domain EC5 in homophilic adhesion. EC5 might not participate in the adhesive bond, possibly explaining the preservation of homophilic adhesion in EC5-deficient mutants. Alternatively, conformational changes in E-cadherin resulting from the EC5 deletion may cause cell aggregation via a process different from that occurring with the native molecule. However, current data that the same mAbs inhibiting homotypic aggretgates by wild-type transfectants (SHE78-7, HECD-1) also inhibited those by the 5 mutant would be evidence that the process of cell aggregation is similar.
Consistent with previous studies (21, 22, 45), we found that Mn2+ stimulated heterophilic interactions between E7 and E-cadherin. The functional activity of integrins is regulated through an inside-out signaling mechanism that quickly switches inactive forms to active forms. Divalent cations are also required for the acquisition of the active state. In particular, manganese has been shown to promote ligand binding by inducing a conformational change in the metal ion-dependent adhesion site (46, 47, 48).
To examine homophilic and heterophilic interactions, we adopted a cell aggregation assay that has been frequently used to evaluate cell-cell adhesion activity (4, 15, 35, 37). This assay method produced reproducible and clear results in the present study. When 1, 2, 3, or 4 transfectants were mixed with E7+ K562 cells, aggregation did not occur, indicating that all four domain-deletion mutants had lost their ability to undergo heterophilic adhesion with E7. Alternatively, aggregation may have been prevented by a disruption in homophilic adhesion, although the ability to undergo heterophilic adhesion was retained. Heterophilic interactions have been suggested to be so weak that cell aggregation may not occur without homophilic adhesion; alternatively, E7 adhesion may require an E-cadherin homophilic bond to be formed by trans-interactions among cells. The mAb HECD-1 against EC2 did not inhibit homophilic aggregation completely, allowing heterophilic adhesion and supporting the possibility that a homophilic scaffold may be necessary for heterophilic interactions. If so, the effect of the 1, 2, 3, and 4 mutants cannot be evaluated because these mutants prevent homophilic adhesion, and some degree of homophilic adhesion by the transfected L cells is a prerequisite to allowing a determination of the percentage of K562 cells included in the aggregates of E-cadherin-expressing cells. Therefore, the EC5 deletion mutant is the only mutant of the entire set used in this article that can be evaluated for loss of heterophilic adhesion in the K562 coaggregation assay.
We demonstrated that EC5 is critical for heterophilic adhesion with E7+ cells, but not for homophilic adhesion. This is the first evidence showing the involvement of ECs other than EC1 in heterophilic adhesion with E7. Since the integrin E7 plays a critical role in the selective localization of T cells in inflamed epithelia (26, 27, 28, 29, 30, 31, 49), the adhesive interaction could be a potential target for therapeutic interventions. In contrast, mutational analyses of selected residues clearly revealed that the side chain of Glu31 is important for integrin E7 recognition, but not for homophilic adhesion, indicating that the E-cadherin residues critical for heterophilic adhesion to E7 are distinct from those required for homophilic adhesion (32, 33). Higgins et al. (50) proposed a docking model involving the E A domain and EC1 in which the metal ion-dependent adhesion site cleft of E comes in contact with Glu31 of EC1 and the Phe298 projection of E coordinates with the hydrophobic pocket of EC1. Given a previous report showing that synthetic peptides encompassing Asn27-Val34 in EC1 had very little inhibitory effect on the interaction with E7 (32), however, full adhesive activity may require other binding sites or events, such as a conformational change in the cadherin or integrin molecules. Since domain 5 is located proximal to the membrane, it is less likely to serve as a direct binding site for the integrin, implying that the effect of the EC5 deletion on binding of the E7 may be due to long-range effects that alter the conformation of more membrane distal domain of the molecules. However, the 5 mutants retained the ability to participate in homophilic adhesion and still bound the EC1-specific Ab SHE78-7 and the EC2-specific Ab HECD-1, indicating that a gross disruption of the conformation of the domain distal to EC5 is probably not occurring. The complexity of the EC5 conformational structure, containing asparagine residues of N-linked glycosylation and two intramolecular disulfide bonds, may favor the hypothesis that EC5 plays a role in the regulation of heterophilic adhesion via a change in conformation of the membrane proximal domain. Although the inside-out signaling mechanism that alters the conformational change in E-cadherin remains to be clarified, further functional and structural studies should provide an insight into our understanding of the role of EC5 in heterophilic adhesion. Identification of smaller deletions or point mutations in EC5 that also affected heterophilic adhesion by E7 would be of considerable interest. The current data did not show that EC5-specific Ab G-10 block heterophilic adhesion of E7 to E-cadherin. The adhesion was also not inhibited by the polyclonal Ab, which we newly generated in rabbits using a synthetic peptide corresponding to part of EC5 (our unpublished observations). If any anti-EC5 mAb blocks heterophilic adhesion to the same degree as anti-E, this would also be good supporting evidence for a role of EC5 in heterophilic adhesion.
Acknowledgments
We are grateful to Dr. D. Erle for providing K562-E7 cells and Dr. Y. Shimoyama for providing the vector containing E-cadherin cDNA. We acknowledge the use of equipment belonging to the Saitama Medical School Research Center for Genomic Medicine
Disclosures
The authors have no financial conflict of interest.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 Address correspondence to Dr. Kiyono Shiraishi, Project Research Laboratory, Research Center for Genomic Medicine, Saitama Medical School, 1397-1 Yamane, Hidaka, Saitama 350-1241, Japan. E-mail address: kiyono@saitama-med.ac.jp
2 Abbreviations used in this paper: EC, extracellular domain; DiO, 3,3'-deoctadecyl-5, 5'-di(4-sulfophenyl)oxacarbocyanine sodium salt.
Received for publication January 5, 2005. Accepted for publication May 13, 2005.
References
Takeichi, M.. 1991. Cadherin cell adhesion receptors as a morphogenetic regulator. Science 251: 1451-1455.
Takeichi, M.. 1995. Morphogenetic roles of classical cadherins. Curr. Opin. Cell Biol. 7: 619-627.
Yap, A. S., W. M. Brieher, B. M. Gumbiner. 1997. Molecular and functional analysis of cadherin-based adherens junctions. Annu. Rev. Cell Dev. Biol. 13: 119-146.
Nose, A., K. Tsuji, M. Takeichi. 1990. Localization of specificity determining sites in cadherin cell adhesion molecules. Cell 61: 147-155.
Overduin, M., T. S. Harvey, S. Bagby, K. I. Tong, P. Yau, M. Takeichi, M. Ikura. 1995. Solution structure of the epithelial cadherin domain responsible for selective cell adhesion. Science 267: 386-389
A. Thompson, M. S. Lehmann, G. Grübel, J. F. Legrand, J. Als-Nielsen, D. R. Colman, W. A. Hendrickson. 1995. Structural basis of cell-cell adhesion by cadherins. Nature 374: 327-337.(Kiyono Shiraishi1,*,, Ken)
The integrin E7 is expressed on intestinal intraepithelial T lymphocytes and CD8+ T lymphocytes in inflammatory lesions near epithelial cells. Adhesion between E7+ T and epithelial cells is mediated by the adhesive interaction of E7 and E-cadherin; this interaction plays a key role in the damage of target epithelia. To explore the structure-function relationship of the heterophilic adhesive interaction between E-cadherin and E7, we performed cell aggregation assays using L cells transfected with an extracellular domain-deletion mutant of E-cadherin. In homophilic adhesion assays, L cells transfected with wild-type or a domain 5-deficient mutant formed aggregates, whereas transfectants with domain 1-, 2-, 3-, or 4-deficient mutants did not. These results indicate that not only domain 1, but domains 2, 3, and 4 are involved in homophilic adhesion. When E7+ K562 cells were incubated with L cells expressing the wild type, 23% of the resulting cell aggregates consisted of E7+ K562 cells. In contrast, the binding of E7+ K562 cells to L cells expressing a domain 5-deficient mutant was significantly decreased, with E7+ K562 cells accounting for only 4% of the cell aggregates, while homophilic adhesion was completely preserved. These results suggest that domain 5 is involved in heterophilic adhesion with E7, but not in homophilic adhesion, leading to the hypothesis that the fifth domain of E-cadherin may play a critical role in the regulation of heterophilic adhesion to E7 and may be a potential target for treatments altering the adhesion of E7+ T cells to epithelial cells in inflammatory epithelial diseases.
Introduction
E-cadherin, a classic member of the cadherin superfamily, is expressed on epithelial cells and mediates Ca2+-dependent homophilic cell-cell adhesion (1, 2, 3). Classic cadherins contain five extracellular domains (ECs)2 of 110 aa each, a transmembrane domain, and a cytoplasmic domain. Structure-function analyses of the homophilic interactions of E-cadherin have largely focused on the NH2-terminal EC domain (EC1), which contains a highly conserved His-Ala-Val motif (4, 5, 6, 7). Indeed, protein fragments or peptides containing the His-Ala-Val sequence exerted limited effects on cell-cell adhesion (8, 9). Recently, crystallographic analysis has clearly demonstrated that conserved Trp in the EC1 domains of classical cadherins is critical for trans-interactions between E-cadherin molecules on different cells, serving as a strand dimer (10). However, several studies have suggested that ECs may be involved in cell adhesion in ways other than the role mediated by EC1. In human cancers, for example, E-cadherin gene mutations frequently occur in exons 7, 8, and/or 9 (corresponding to EC2 and EC3); these mutations are thought to result in the loss of the ability to undergo cell-cell adhesion (11, 12, 13, 14). A study on the binding properties of the soluble C-cadherin ectodomain suggested that EC1 was not sufficient for complete homophilic binding (15). Furthermore, Corada et al. (16, 17) demonstrated that mAbs directed against EC3-EC4 affected VE-cadherin adhesion in endothelial cells.
The heterophilic interaction of E-cadherin and integrin E7 has been previously documented (18, 19, 20, 21, 22, 23, 24). Integrin E7 is expressed selectively on intestinal intraepithelial T lymphocytes under physiological conditions (25). Accumulating evidence indicates that E7+ is induced on T lymphocytes in the epithelia of skin, lung, salivary, and lacrimal glands and synovial membranes during inflammation (26, 27, 28, 29, 30, 31), suggesting that this heterophilic interaction has a pathologic role. Since E7 may have an important role in the selective localization or retention of a unique population of T cells in a specific tissue, the adhesion between E7 and E-cadherin could be a potential target of therapeutic interventions for epithelial inflammation. Substitutions of a highly exposed and charged amino acid on mouse E-cadherin transfected into L cells demonstrated that Glu31 in EC1 is critical for binding with E7 (32). Taraszka et al. (33) also elucidated that the substitution of the corresponding Glu in EC1 of human E-cadherin-Fc fusion proteins abrogated binding with E7, confirming the previous observations. These studies clearly show that the specific heterophilic adhesion attributed to the exposed Glu31 in EC1 differs from homophilic adhesion. With the exception of EC1, however, the structures involved in heterophilic adhesion remain uncertain. To clarify the involvement of the EC domains in both heterophilic and homophilic adhesion, we performed cell aggregation assays using L cells transfected with specific domain-deleted E-cadherin mutations. The present report speculates on the characteristics of the E-cadherin domains involved in homophilic interactions and heterophilic interactions with E7.
Materials and Methods
Following the cell aggregation assays, the cell aggregates were collected and washed with PBS. For immunofluorescence staining, the aggregates were incubated with anti-E-cadherin Ab (HECD-1). Subsequently, the samples were incubated with Alexa Fluor 568-conjugated goat anti-mouse IgG Ab (Molecular Probes), followed by incubation with FITC-conjugated anti-E7 Ab (Beckman Coulter). The specimens were observed using a confocal laser-scanning microscope.
Results
To confirm whether the transfectants expressed the desired region of E-cadherin, RT-PCR was performed for EC1, EC2, EC3, EC4, and EC5. Fig. 2 shows that the parent L cells did not exhibit any products amplified by human E-cadherin and that all five ECs were expressed in wild-type transfectants. All clones transfected with domain-deletion mutations of E-cadherin exhibited the expected expression patterns: 1 lacked EC1, 2 lacked EC2, 3 lacked EC3, 4 lacked EC4, and 5 lacked EC5.
The protein expression of mutated E-cadherin was examined by Western blot analysis using the appropriate Abs (Fig. 3a). 4A2C7, raised against a recombinant protein corresponding to the cytoplasmic domain, detected all of the deletion mutants. G-10, an alternative E-cadherin Ab that cross-reacts with mouse form (manufacturer’s data), could not detect endogenous mouse E-cadherin in mock transfectants (data not shown). To clarify the binding epitopes recognized by the SHE78-7, HECD-1, and G-10 Abs, GST fusion proteins containing the EC domains were analyzed (Fig. 3b). SHE78-7 reacted with EC1, whereas HECD-1 reacted with EC2. G-10, raised against a recombinant protein corresponding to EC5, correctly detected EC5.
Discussion
In this study, we attempted to clarify the overall binding capability of E-cadherin for both homophilic adhesion and heterophilic adhesion to the integrin E7 through the generation of domain-deletion E-cadherin mutants and the transfection of these mutants into L cells to determine the roles of the individual domains in cell adhesion. EC1, EC2, EC3, and EC4 domain-deletion mutants lost their homophilic binding ability, suggesting that these EC domains are indispensable for homophilic adhesion. Substantial evidence suggests that the combination of cis-dimerization of two cadherin molecules on the same cell surface and trans-interactions between cadherin dimers on opposing cell surfaces maximizes homophilic adhesion. The widely accepted linear zipper model attributes the adhesive interfaces in the cis- and trans-interactions to EC1 (6, 7, 36). However, recent studies have introduced a new model for homophilic adhesion. Through the analysis of domain deletions in the Xenopus C-cadherin ectodomain using bead aggregation and cell adhesion assays, Clappuis-Flament et al. (15) demonstrated that the combination of at least three EC domains, such as EC1-EC2-EC3 or EC1-EC2-EC4, was required for trans-interaction and proposed an alternative model in which multiple ECs are required to achieve full adhesive capability. Several studies have provided data to support this model, suggesting the existence of multiple adhesive interfaces. For example, a mAb recognizing a potential epitope spanning the EC3–EC4 domains affected VE-cadherin adhesion in endothelial cells from umbilical veins (16). A mutant lacking the EC1 domain failed to exhibit homophilic adhesion, although the mutant could adhere to other mutants expressing wild-type E-cadherin (37). In contrast, a crystallographic analysis of C-cadherin supported the linear zipper model by showing that EC1 interacts with EC2 on other E-cadherin molecules on the same cell surface, leading to cis-dimerization (10). Our results showing that EC1- or EC2-deficient mutants failed to exhibit full homophilic adhesion are basically consistent with the above findings. However, mutants lacking either the EC3 or EC4 domains also failed to exhibit maximum adhesion. The fact that the present study used mammalian cadherin, whereas the previous study used Xenopus cadherin, may partially explain this discrepancy in findings. Another possible explanation may be that intracellular events were affected by the domain deletion in our aggregation assay, since the transmembrane and cytoplasmic domains were included in the domain-deletion human E-cadherin mutants. Compelling evidence suggests that the adhesive strength of cadherin is regulated by the transmembrane and cytoplasmic domains (3, 38, 39, 40, 41, 42). For example, the cytoplasmic tail of cadherin interacts with -catenin and p120, which link cadherin to the actin cytoskeleton through -catenin, and the regulation of the cadherin-catenin complex by diverse phosphorylation reactions influences the adhesive function of cadherin (43). In addition, the interaction of a motif in domain 4 of N-cadherin with the fibroblast growth factor receptor is required for neurite outgrowth (44), suggesting that EC4 may be involved in a cell signaling pathway in which fibroblast growth factor receptor controls the gene transcription and adhesive activity of cadherin via Snail, resulting in a loss of cell-cell adhesion (43). To the best of our knowledge, little information is available about the involvement of the membrane proximal domain EC5 in homophilic adhesion. EC5 might not participate in the adhesive bond, possibly explaining the preservation of homophilic adhesion in EC5-deficient mutants. Alternatively, conformational changes in E-cadherin resulting from the EC5 deletion may cause cell aggregation via a process different from that occurring with the native molecule. However, current data that the same mAbs inhibiting homotypic aggretgates by wild-type transfectants (SHE78-7, HECD-1) also inhibited those by the 5 mutant would be evidence that the process of cell aggregation is similar.
Consistent with previous studies (21, 22, 45), we found that Mn2+ stimulated heterophilic interactions between E7 and E-cadherin. The functional activity of integrins is regulated through an inside-out signaling mechanism that quickly switches inactive forms to active forms. Divalent cations are also required for the acquisition of the active state. In particular, manganese has been shown to promote ligand binding by inducing a conformational change in the metal ion-dependent adhesion site (46, 47, 48).
To examine homophilic and heterophilic interactions, we adopted a cell aggregation assay that has been frequently used to evaluate cell-cell adhesion activity (4, 15, 35, 37). This assay method produced reproducible and clear results in the present study. When 1, 2, 3, or 4 transfectants were mixed with E7+ K562 cells, aggregation did not occur, indicating that all four domain-deletion mutants had lost their ability to undergo heterophilic adhesion with E7. Alternatively, aggregation may have been prevented by a disruption in homophilic adhesion, although the ability to undergo heterophilic adhesion was retained. Heterophilic interactions have been suggested to be so weak that cell aggregation may not occur without homophilic adhesion; alternatively, E7 adhesion may require an E-cadherin homophilic bond to be formed by trans-interactions among cells. The mAb HECD-1 against EC2 did not inhibit homophilic aggregation completely, allowing heterophilic adhesion and supporting the possibility that a homophilic scaffold may be necessary for heterophilic interactions. If so, the effect of the 1, 2, 3, and 4 mutants cannot be evaluated because these mutants prevent homophilic adhesion, and some degree of homophilic adhesion by the transfected L cells is a prerequisite to allowing a determination of the percentage of K562 cells included in the aggregates of E-cadherin-expressing cells. Therefore, the EC5 deletion mutant is the only mutant of the entire set used in this article that can be evaluated for loss of heterophilic adhesion in the K562 coaggregation assay.
We demonstrated that EC5 is critical for heterophilic adhesion with E7+ cells, but not for homophilic adhesion. This is the first evidence showing the involvement of ECs other than EC1 in heterophilic adhesion with E7. Since the integrin E7 plays a critical role in the selective localization of T cells in inflamed epithelia (26, 27, 28, 29, 30, 31, 49), the adhesive interaction could be a potential target for therapeutic interventions. In contrast, mutational analyses of selected residues clearly revealed that the side chain of Glu31 is important for integrin E7 recognition, but not for homophilic adhesion, indicating that the E-cadherin residues critical for heterophilic adhesion to E7 are distinct from those required for homophilic adhesion (32, 33). Higgins et al. (50) proposed a docking model involving the E A domain and EC1 in which the metal ion-dependent adhesion site cleft of E comes in contact with Glu31 of EC1 and the Phe298 projection of E coordinates with the hydrophobic pocket of EC1. Given a previous report showing that synthetic peptides encompassing Asn27-Val34 in EC1 had very little inhibitory effect on the interaction with E7 (32), however, full adhesive activity may require other binding sites or events, such as a conformational change in the cadherin or integrin molecules. Since domain 5 is located proximal to the membrane, it is less likely to serve as a direct binding site for the integrin, implying that the effect of the EC5 deletion on binding of the E7 may be due to long-range effects that alter the conformation of more membrane distal domain of the molecules. However, the 5 mutants retained the ability to participate in homophilic adhesion and still bound the EC1-specific Ab SHE78-7 and the EC2-specific Ab HECD-1, indicating that a gross disruption of the conformation of the domain distal to EC5 is probably not occurring. The complexity of the EC5 conformational structure, containing asparagine residues of N-linked glycosylation and two intramolecular disulfide bonds, may favor the hypothesis that EC5 plays a role in the regulation of heterophilic adhesion via a change in conformation of the membrane proximal domain. Although the inside-out signaling mechanism that alters the conformational change in E-cadherin remains to be clarified, further functional and structural studies should provide an insight into our understanding of the role of EC5 in heterophilic adhesion. Identification of smaller deletions or point mutations in EC5 that also affected heterophilic adhesion by E7 would be of considerable interest. The current data did not show that EC5-specific Ab G-10 block heterophilic adhesion of E7 to E-cadherin. The adhesion was also not inhibited by the polyclonal Ab, which we newly generated in rabbits using a synthetic peptide corresponding to part of EC5 (our unpublished observations). If any anti-EC5 mAb blocks heterophilic adhesion to the same degree as anti-E, this would also be good supporting evidence for a role of EC5 in heterophilic adhesion.
Acknowledgments
We are grateful to Dr. D. Erle for providing K562-E7 cells and Dr. Y. Shimoyama for providing the vector containing E-cadherin cDNA. We acknowledge the use of equipment belonging to the Saitama Medical School Research Center for Genomic Medicine
Disclosures
The authors have no financial conflict of interest.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 Address correspondence to Dr. Kiyono Shiraishi, Project Research Laboratory, Research Center for Genomic Medicine, Saitama Medical School, 1397-1 Yamane, Hidaka, Saitama 350-1241, Japan. E-mail address: kiyono@saitama-med.ac.jp
2 Abbreviations used in this paper: EC, extracellular domain; DiO, 3,3'-deoctadecyl-5, 5'-di(4-sulfophenyl)oxacarbocyanine sodium salt.
Received for publication January 5, 2005. Accepted for publication May 13, 2005.
References
Takeichi, M.. 1991. Cadherin cell adhesion receptors as a morphogenetic regulator. Science 251: 1451-1455.
Takeichi, M.. 1995. Morphogenetic roles of classical cadherins. Curr. Opin. Cell Biol. 7: 619-627.
Yap, A. S., W. M. Brieher, B. M. Gumbiner. 1997. Molecular and functional analysis of cadherin-based adherens junctions. Annu. Rev. Cell Dev. Biol. 13: 119-146.
Nose, A., K. Tsuji, M. Takeichi. 1990. Localization of specificity determining sites in cadherin cell adhesion molecules. Cell 61: 147-155.
Overduin, M., T. S. Harvey, S. Bagby, K. I. Tong, P. Yau, M. Takeichi, M. Ikura. 1995. Solution structure of the epithelial cadherin domain responsible for selective cell adhesion. Science 267: 386-389
A. Thompson, M. S. Lehmann, G. Grübel, J. F. Legrand, J. Als-Nielsen, D. R. Colman, W. A. Hendrickson. 1995. Structural basis of cell-cell adhesion by cadherins. Nature 374: 327-337.(Kiyono Shiraishi1,*,, Ken)